Relative Density vs Relative Compaction and in-situ testing

[novicesoilserf]

I have a project site with 30-100 feet of non-engineered fill. I've conducted hollow-stem soil borings and have collected samples with SPT and California Modified samplers. The blow counts were recorded and corrected to (N1)60 blow counts with an additional correction for the Cal. Mod. samples. I reported the results as relative densities. I am now being told that I need to determine the relative compaction of the various layers by getting the lab dry densities of the rings and SPT tubes and report their percentages of the maximum dry densities (as determined by ASTM D 1557). The source of this request went on to say that blow counts should be used in support of these lab testing comparisions and not the other way around.

I've always thought relatively disturbed/undisturbed samples should not be used to determine relative compaction, especially with this much varied overburden. Could someone please clarify?

"Information is not knowledge" Albert Einstein

 

[Ron]

The request is not practical. First of all, you probably have limited sample available. To run a proper moisture-density relationship in the lab (ASTM D1557) you will need about 30 lbs of soil. You won't get that from a split spoon or even a Shelby tube, unless you take a lot of them. Then you have to determine if there is any consistency to the soil classification if you are doing a composite sample.

IMHO, you have the field data. Use correlations as best you can...you're chasing an accuracy that you won't find. 30 to 100 feet of fill without any control is not a good thing. Probably the best remediation you can provide is to recommend surcharging and monitor the settlement of the surcharge.

 

[novicesoilserf]

Thanks Ron!

To add some information, I did collect some bulk samples, but they were of course collected from the hollow stem cuttings. The cuttings are pretty representative of discreet sampled materials at shallow depths, but as the bulks are collected from deeper depths (>10 feet) the soil is being mixed from the cutting head all the way to the ground surface. Since I only have "windows" of observation at 5-foot intervals, I cannot say for sure what is happening between these windows (engineering judgement at it's finest). To be honest, the material throughout the depths is a pretty uniform brown silty sand with little (<15%) to no clays or gravels.

My big question still remains, so I will rephrase it. Does testing of discreet relatively undisturbed samples (i.e. Cal Mod sampler rings) truly give reasonable relative compaction values? Is this method better (or more reliable) than using the established SPT correlations to relative density? Ok, so I guess technically that's two questions still remaining .

"Information is not knowledge" Albert Einstein

 

[moe333]

Even if you could find a way to obtain enough "representative" material to run ASTM D1557 the modified Cal is not an undisturbed sample - particularly with a granular material such as SM - the actual dry density will be effected and you won't know by how much. I doubt you could learn anything meaningful from comparing the Cal density with ASTM D1557.

 

[oldestguy]

Ron:

Didn't these forums at least once some time back go over the term "relative compaction" and the majority didn't recognize that term as a commonly used term or calculation?

 

[novicesoilserf]

Thanks oldestguy. I was hoping to get some insight from you, BigH, fattdad, and/or Focht3.

The question still remains though, do relative compaction values(relative undisturbed/disturbed ring sample dry densities as a percentage of lab max dry densities) more accurately define the soil stratigraphy than corrected blow counts (correlated to established relative densities).

Please note that Relative Compaction does not equal Relative Density and these two terms should never be used interchangeably (e.g. 80% relative compaction equals about 0% relative density), see Lee, K.L., and Singh, A. (1971) "Compaction of Granular Soils," Proceedings of the Ninth Annual Symposium on Engineering Geology and Soils Engineering, Boise, Idaho, pp 161-174.

"Information is not knowledge" Albert Einstein

 

[Ron]

OG...you're right. There was a discussion on that.

 

[dgillette]

I never cared much for the term "relative compaction." Sounds too much like "relative density," and the difference cannot be inferred from the words, like street bike/road bike, dirt bike/mountain bike. Which ones have motors?

My usual term is "percent compaction," or "percent of lab max," although the term "D value" is used some places (limited to western US?).

 

[BigH]

I've used "relative compaction" and know the difference in it and "relative density". One needs to understand the context and engineering definitions - not the common vernacular - similar to the arguments I have made with allowable bearing capacity (based on shear) and allowable bearing pressure (based on serviceability).

I think it a bit disingenuous of the person asking the OP to estimate the relative compaction from the N values. The N values are reflective of many things such as angularity of the particles, particle size vs the spoon's diameter and the like and this doesn't always translate to a general correlation.

The question is whether this is to show that there was no control on the fill being placed (a contractual thing) or the desire to build on the land (an engineering thing). if the former, one is probably way out of luck except in a qualitative sense that shows a very large COV for the N values obtained. This would indicate uncontrolled fill. If for the engineering, one would look at it in the sense of any soils investigation and apply design principals accordingly fully aware, though, of the large variations that are likely (but not necessarily) inherent in an uncontrolled fill.

 

[fattdad]

If the ultimate goal is to ascertain whether you can put a building on the old fill, there are other ways. . .

First I agree that the California sampler (i.e., with the brass rings, etc.) is not undisturbed. While you can get a density, it will be somewhat approximate. Further what maximum dry density will you use for comparison - some surface-obtained bulk sample? Likley to be just another approximation.

Let's say you use all these approximations and "calculate" some value like 92 percent relative compaction. So what? I mean would that mean you can't build on the stuff? Would that mean you'd have to use piles? No it wouldn't.

I'd jam a dilatometer or a cone in the ground and directly measure the material properties and assess the settlement/stability potential using those data. I don't tell you that as some glib response. I tell you this 'cause that's what I'd do and have done in the past. After all, the relative compaction is moot if you can show some measure of subbase improvement in conjunction with the in-situ properties of the fill are such that the structural loads and foundations can function for the intended development.

f-d (back from vacation)

¡papá gordo ain’t no madre flaca!

 

[molerat2210]

I don't see anything wrong with collecting tube samples from a fill, if there is little gravel. It is not perfectly accurate but what is? It can give you an idea of the density. The density of the tube samples can be compared to the proctor of a bulk sample collected from a similar depth.

A few test pits can be excavated, if space allows. Nuke density tests can be performed at various depths to about 10 feet to compare with a few tube samples. I like test pits when using an existing fill for some engineering purpose to make sure there are few surprises like rubble and organic matter although I understand 100 feet of fill is extreme. Maybe a deep exploratory shaft is warranted with incremental density tests?

Drive cylinder samplers can be used at the bottom of the test pit instead of Shelby tubes if it's clayey enough.

CPTs are also a good way to collect lots of data from a fill that can be used with Schmertmann's method to calculate settlement.

Correlations of density with SPTs is a little weak. One recent dinky structure we worked on was on a rubble fill of blast rock. It needed 400 cy of grout and a mat slab to get things under control. The SPTS were high from the rubble but full of voids. The voids were narrow seams between boulders and not well shown by the borings. Test pits painted the picture more clearly.

 

[novicesoilserf]

Thanks for all the input. Wish me luck.

"Information is not knowledge" Albert Einstein

 

[aeoliantexan]

The disturbance of the California samples depends on the details of the sampler, among other things. A sharp and slightly undersized cutting tip helps reduce disturbance.

If the soil is unsaturated,the disturbance will tend to be densification and strengthening of the sample, therefore unconservative.

Whatever the properties of this soil are today, they can be drastically changed by future increases in water content. Re-grading, irrigation, and climatic rainfall cycles can cause big changes in strength and compressibility. Even a well-compacted fill that thick can compress 1 or 2 percent when wetted.

I would be very cautious about predicting the long-term performance of anything supported on this fill, even after preloading.